OpenZFS, a robust, open-source file system and volume manager that integrates advanced data protection features like snapshots, checksums, and replication, has just released its latest version, 2.3.
One of the standout features is RAIDZ Expansion, which allows users to add new devices to existing RAIDZ pools to increase storage capacity seamlessly without any downtime.
In addition, a major upgrade to Fast Dedup has been introduced to improve deduplication performance, which is sure to catch the attention of anyone looking to optimize storage efficiency.
Furthermore, Direct IO lets you bypass the Adaptive Replacement Cache (ARC) for reads and writes—this is especially handy when you have speedy NVMe drives that might be slowed down by caching.
The introduction of JSON output provides users with a more standardized way of parsing essential system commands, and support for long file and directory names now extends up to a whopping 1023 characters.
On top of that, the release addresses a series of critical bug fixes, incorporates numerous performance improvements, and remains compatible with Linux kernels 4.18 through 6.12 as well as FreeBSD versions 13.3 and 14.0–14.2.
Among the new module parameters are “dmu_ddt_copies,” “spa_num_allocators,” “zfs_vdev_direct_write_verify,” and many others that can be tuned to suit specific performance goals or operational requirements.
Notably, “zfetch_array_rd_sz” was removed, while options such as “zfs_arc_shrinker_limit” and “zfs_vdev_disk_classic” have been modified. If you want to dig deeper into each one, you can check out the official module parameters documentation for a full breakdown.
Lastly, the OpenZFS team has thoughtfully curated detailed documentation and a change log covering the jump from v2.2 to v2.3, making it easy to stay informed about every tweak under the hood.
For more information, see the changelog.
